On Mon, 2020-02-24 at 13:33 -0800, Brian Norris wrote: > > > if not, IMHO better to add some > > advertisement bits, but if not then we can surely add some kind of > > filter callback that's invoked at config time, rather than only at > > suspend time when it's way too late to do anything about it. > > Yeah, I'd have to see what sort of restrictions we're working with -- > sufficiently complex rules might not make for good "advertisement" > bits. Agree. And we've certainly gotten better at spelling out what exactly when wrong since all this started - the "extended ACK" bits I wrote that can include an error message and can point to the bad attribute still feel sort of new to me, even if we've had them for a few years :) So if we end up going with a callback, we should probably include an extack pointer for the driver to fill up ... johannes
